4 seconds of pace. What to train? :-)

Discussion in 'General Discussion' started by Nygaard, Oct 13, 2017.

  1. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,425
    Hi. Let me explain.

    First of all you pointed a good point that I didn't test. Braking. I will add a straight line braking test to the set.

    All the tests are done with keyboard controls in order to be repetitive. For maintaining steering angle I use full lock while lowering the steering lock range setting to minimum in order to minimize spins. Acceleration from zero in this case is done at a high gear ratio for the same reason. Different starting gears can be tested this way for different acceleration levels.

    Maintaining constant steering does not provoke a constant radius turn!!! You are very wrong in that. In fact I have plotted turning radius in the images above. I forgot to build the track layout using that stint in order to see the spiral pattern. Turning radius increases with speed.

    I disagree on your statement that the car has low power to compare acceleration as representative: It is the way to do it and not even the slightest difference has been seen. The conclusion is clear in terms of power they behave the same. Without having done the braking tests it is quite evident that the result will be the same.

    If you would prefer another car to test let me know.
     
  2. Nygaard

    Nygaard Registered

    Joined:
    Feb 1, 2017
    Messages:
    29
    Likes Received:
    8
    I may be wrong on this one but as i remember(lack of memory can occur :)) you can pretty much floor this car with traction control off, so except from the very start traction or grip is not an issue when accelerating in a straight line from power. So i fail to see why different grip will produce different results unless the difference is so big that traction or grip will become the limiting factor.

    But in your case the speed is almost constant in the second half, So i assumed the turning radius was the same. (cant see turning radius in figure). Okay you've actually build a new track layout for this test. Cool.

    I am aware this will take some time and work :), but:
    If it was possible i think the most effiecient way to test if grip/traction is different from online and offline session would be to actually draw circles on the track. In the test a constant radius could be manually controlled by keeping the car at the drawn circle. The test in which the car skid at the highest speed has the most grip.

    Does the car start to skid in your test? If not, how do you determine if there is any difference in grip/traction
     
    SPASKIS likes this.
  3. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,425
    Losing Grip is actually tested in those pictures. There is overall wheel spin in the fist moments of straight line acceleration test when using first gear.

    In turning test, wheels have local sliding and cause variation of slip angles in all wheels. You can see with motec that they are the same. Those slip angles vary with speed being the reason why actually turning radius changes with speed as well.

    I will repeat the tests with another car (please tell me which one do you think). For that one I will also do a constant speed turning test. In order to properly have control of speed I will provoke a sharp torque drop at the max power rpm.
    Changing gear ratios I will be able to adjust those rpm to the desired speed.
    For this reason I cannot use encrypted cars like just released GT3.
     
  4. Nygaard

    Nygaard Registered

    Joined:
    Feb 1, 2017
    Messages:
    29
    Likes Received:
    8
    . Ok I now see that at 1500 meters, and I can see that the car rotates more as well and the car stop increasing the speed even though full throttle is replied. Is this the same for both cases offline and online?
     
  5. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,425
    Yes, both cases provide overlapping plots
     
  6. Nygaard

    Nygaard Registered

    Joined:
    Feb 1, 2017
    Messages:
    29
    Likes Received:
    8
    Ohh.. I now se the black lines showing the other case on top of the other. I just realized that now. :)
     
    SPASKIS likes this.
  7. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,425
    upload_2017-10-19_22-30-12.png upload_2017-10-19_22-39-28.png
    I updated images to show spiral path.
    Cursor is now in the grip losing area.
     

    Attached Files:

  8. SPASKIS

    SPASKIS Registered

    Joined:
    Sep 7, 2011
    Messages:
    3,155
    Likes Received:
    1,425
    I repeated the tests because I had forgotten to eliminate speed steering sensistivity coming with default keyboard settings.
    I realized since the expected increase in spiral radius should be more progressive!
    Despite being at zero it can be perceived that Tire Toe still decreases from -14º to real steered angle (7º). At 70 kph it stabilizes in 7º. Interesting.
    This suggests some different tire model being used at lower speeds. The behavior of all parameters show a trend changeat that speed.
    The continuity of all plots suggests that some damping (canned) factor is being introduced at lower speeds and graduately being eliminated until reaching 70 kph speed at which it is fully removed. ¿?¿?

    I also included a "brake test" at the end of the straight line acceleration test. I had to offset graphs a little bit to match both braking moments
    As in previous tests, not a single difference between Offline (Eduardo Albaizar) and Online (Spaskis): Graphs overlap.

    upload_2017-10-19_23-26-52.png
    upload_2017-10-19_23-14-46.png
     
    Nygaard likes this.
  9. Nygaard

    Nygaard Registered

    Joined:
    Feb 1, 2017
    Messages:
    29
    Likes Received:
    8
    I need to play with this also, when I get the time... :)
     
  10. Emery

    Emery Registered

    Joined:
    Oct 24, 2010
    Messages:
    2,864
    Likes Received:
    1,475
    In the Pacejka magic formula (e.g. rF1), you have to switch to a different model due to divide by zero problem at v=0. I'm not sure how the brush model (rF2) blends into the magic formula when it comes to these calcs and whether it avoided divide by zero, but you may have found evidence that brush model doesn't avoid the math problem?
     
  11. The Iron Wolf

    The Iron Wolf Registered

    Joined:
    Feb 20, 2016
    Messages:
    992
    Likes Received:
    983
    Curious, how is collision model different?
     
  12. Emery

    Emery Registered

    Joined:
    Oct 24, 2010
    Messages:
    2,864
    Likes Received:
    1,475
    Online has to consider prediction of path rather than actual path due to ping. That gets reconciled before collision is applied and provides some leeway.
     
    Nygaard, Lgel, SPASKIS and 1 other person like this.
  13. Nygaard

    Nygaard Registered

    Joined:
    Feb 1, 2017
    Messages:
    29
    Likes Received:
    8
    Conclusion: I can now only blame my self for being slow ;-)
     
    Emery and SPASKIS like this.
  14. Led566

    Led566 Registered

    Joined:
    Jan 25, 2012
    Messages:
    358
    Likes Received:
    24
  15. davehenrie

    davehenrie Registered

    Joined:
    Jul 6, 2016
    Messages:
    6,019
    Likes Received:
    3,500
    That looks like a conversion that did not match the exact dimensions Simtek was using. The wheelbase is too long. For GTR2, there was a way to ignore that and either use the model or the dimensions provided by the modder. I have no idea if that is still buried in the rf2 hdc files.
     

Share This Page